<<

Earth and Planetary Science Letters 241 (2006) 594–602 www.elsevier.com/locate/epsl

Weathering of as an climatic indicator

Norman H. Sleep a,*, Angela M. Hessler b

a Department of Geophysics, Stanford University, Stanford, CA 94305, USA b Department of Geology, Grand Valley State University, Allendale, MI 49401, USA Received 25 July 2005; received in revised form 4 November 2005; accepted 11 November 2005 Available online 13 December 2005 Editor: K. Farley

Abstract

Chert and other hard monomineralic quartz grains weather mostly by mechanical processes in modern environments. Their clasts are overrepresented in conglomerates and relative to their sources regions. Conversely, macroscopic dissolution features, including , are rare but not nonexistent. The similar rarity of quartz dissolution in Archean deposits provides a paleothermometer for climate on the early Earth. For example, is overrepresented in conglomerates and sands of the ~3.2 Ga Moodies Group (South Africa) relative to the source region. Features related to the far-from-equilibrium dissolution rate are particularly diagnostic as it increases an order of magnitude over 25 8C, much more than solubility. Extrapolating from observed dissolution rates in modern environments that weather at ~25 8C, we expect obvious dissolution features in ancient climates above ~50 8C. Polycrystalline quartz and chert would readily disaggregate by solution along grain boundaries, yielding silt and clay. Quartz grains within slowly granite would become friable, yielding silt and clay, rather than . At still higher temperatures, Al2O3-rich clays from weathered granite would stand above solution- weathered chert on low-relief surfaces. The observed lack of these features is evidence that the Archean climate was not especially hot. D 2005 Elsevier B.V. All rights reserved.

Keywords: chert; quartz sand; conglomerate; dissolution; karst; kinetics; climate;

1. Introduction See the work of Young et al. [4] for a possible exception at 2.9 Ga. Climate during the Archean Era (~3.8 to 2.5 Ga) is Theoretical studies of Archean temperature model a relevant to the history of life on Earth. However, only greenhouse atmosphere composed of carbon dioxide limited geological analysis bears on the topic. Oxygen- and methane. The commonly cited upper limit, 85 8C, isotope studies of provide oceanic temperature is model dependent [5–8]. Conversely, reasonable estimates for 3.5–3.2 Ga of 55–85 8C [1]. Evidence of amounts of these gases do not necessarily imply tem- thorough weathering in the Archean, despite the absence peratures much higher than those of the modern tropics of vascular plants, has prompted temperature estimates or the Earth [8]. Limited direct estimates of as high as 85 8C [2,3]. In addition, the limited record atmospheric CO2 come from observations that siderite shows that glaciation was uncommon in the Archean. is present in 3.2-Ga pebble weathering rinds [9] but is absent from a 2.75-Ga paleosol [10] (Fig. 1). * Corresponding author. In this paper, we suggest that the weathering beha- E-mail address: [email protected] (N.H. Sleep). vior of quartz in monomineralic quartzite and chert and

0012-821X/$ - see front matter D 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.epsl.2005.11.020 N.H. Sleep, A.M. Hessler / Earth and Planetary Science Letters 241 (2006) 594–602 595

the mechanically derived conglomerates or poorly frac- tionated volcaniclastic deposits underlying the Moodies Group.

2. Archean geological observations

Hessler studied well-preserved 3.2-Ga clastic sedi- ments from Barberton Mountain Land for her thesis [12]. We concentrate on Moodies Group rocks north of the Inyoka Fault, which have been the subject of ex- tensive geological work [13–17]. A striking feature of these rocks is that the degree of weathering prior to fluvial deposition is within the range of that observed at the warm, humid end of the climate spectrum on the modern Earth. This observation is elaborated below for each sedimentary size fraction within the Moodies Group. We cite other analogous localities from the Mesoarchean and Neoarchean. Fig. 1. Stability diagram of siderite in terms of weathering tempera- ture and the partial pressure of CO2 from Rye et al. [10]. The typical 2.1. Conglomerate weathering temperature is ~10 K hotter than the global mean annual temperature [24]. Our 3.2-Ga pebbles [9] weathered under conditions on the upper left of the diagram while the paleosol of Rye et al. [10] The conglomerate data are straightforward (Fig. 2). weathered at 2.75 Ga under conditions on the lower left. Both situa- At the bottom of the section, interpreted to be alluvial tions are consistent with clement Archean climate. The present (pre- fan and braided river deposits [13,16], conglomerate À 6 industrial) CO2 pressure is ~300Â10 bar. clasts are poorly sorted, spherical, mostly well rounded, and have an average diameter of ~3 cm [9,12]. Chert as grains in granitic rocks is a viable Archean clasts of various colors dominate the pebble fraction of paleothermometer. The tendency on the modern Earth the rock. The rounded clasts upon visual inspection do for quartz to weather mechanically, unlike the dissolu- not have embayments as expected from dissolution. tion weathering of carbonates, is notorious. bAs unre- Although the conglomerates are locally variable, chert lenting to drops of rain.Q (Shakespeare, Titus clasts dominate across the study area (42–58%), despite Andronicus, Act II, Scene III). bHe plies her hard; chert covering b2% of the source region. Other rocks and much rain wears the .Q (Shakespeare, The types that are expected to resist chemical weathering are Third Part of King Henry the Sixth, Act III, Scene II). also overrepresented compared to the source region. There is an urge to regard evidence of this difference on These include quartz-rich and sericite . the ancient Earth as unremarkable. Rather, quartz does Quartz-poor sandstone, ultramafic and mafic volcanic dissolve some in modern environments [11]. Its solu- rocks, and are underrepresented. All the under- bility and dissolution rate of quartz are strong functions of temperature. Quartz will dissolve and chemically disaggregate readily in a sufficiently hot climate, as has been suggested for the Archean. Instead, field observations indicate that quartz in the Archean mostly resisted chemical weathering, like it does under modern clement conditions. Here, we quantify the effect using solubility and kinetic data to extrapolate from the chem- ical behavior of quartz observed in modern climates. Because the Barberton sequence (like most Mesoarchean sequences) is predominately syntectonic, we restrict our attention to derived during Fig. 2. Drill core from the Sheba Gold mine in the Barberton area shows the abundance of chert pebbles (marked C) in a 3.2-Ga periods of relatively slow exhumation, such that chem- Moodies Group conglomerate. The red arrow points to a well-rounded ical weathering could proceed. We eschew stratigraphic mechanically weathered clast. Fractures postdate deposition. U. S. levels dominated by highly immature sediments, like quarter (25 mm) gives scale. 596 N.H. Sleep, A.M. Hessler / Earth and Planetary Science Letters 241 (2006) 594–602 represented rock types, but shale (which is mechanical- nearby Kaap Valley pluton (Fig. 3). Moodies sands ly weak), are expected to weather chemically. are less weathered than those generated in modern, The predominance of chert where it occurs in the extremely humid environments: qualitatively between source region and the mechanical weathering of its the fluvial arkoses of the warm–temperate to subhumid cobbles seem to be a general characteristic of mid- Mallacoota Basin [20] and the quartz–arenite sands Archean sequences (Roger Buick personal communica- produced in the seasonally tropical Orinoco Basin, tion, 2003). An example is 3.5-Ga deposits in Venezuela [21]. [18]. Most of the shale fraction appears to have bypassed the exposed region of deposition. Enough shale exists, 2.2. Sandstone and shale however, to demonstrate that extensive chemical weathering did occur. Al2O3 ranges between 18 and Abundant quartz sandstone in the basal 700–800 m 25 wt.%, above that in any reasonable source rock. of the Moodies Group indicates a braided alluvial plain Conversely, detrital chert comprises less than 2% of [13,16]. Framework grains include abundant monocrys- the shale material. talline quartz (38–39%), potassium (9–15%), Although analogous data from sequences contempo- chert (9–15%), and quartz–sericite lithic grains (6– raneous with the Moodies Group is lacking, extensive 18%), depending on the location within the Barberton chemical weathering is also a common feature of . Less common framework grains in- Mesoarchean and Neoarchean sediments in the Cana- clude plagioclase, polycrystalline quartz, felsic and dian Shield [2,3,22–24] and South Africa [25]. Sand- mafic–ultramafic volcanics, lithic grains, and sedimen- stones in the ~2.6 Ga Beaulieu Formation [3] and the tary lithic grains. The modal composition plots in the ~2.6 Keskarrah Formation [2] show intense chemical recycled orogen field of Dickinson and Suczek [19] weathering in a humid climate, which the authors in- (Fig. 3). It is important to note that chert is overrepre- terpret as being as hot as 85 8C. These samples, how- sented in our sands relative to the source region, but to ever, are more quartz-rich than those in the Moodies a lesser degree than in the conglomerates. Group, plotting in the recycled orogen field of Dick- Overall the degree of weathering is greater than that inson and Suczek [19] at 62–90% [3] and 68–95% [2] observed in sands from modern rivers draining the quartz. As is expected, plagioclase and potassium feld- spar are underrepresented relative to the assumed source region. Notably much of the quartz is polycrys- talline, 85% and 56% of the grains, respectively, in the composite average. Highly mature occur in the ~2.80-Ga Central Slave Cover Group [24]. Howev- er, many of these rocks have recrystallized and point counts of quartz grain types are not available.

2.3. Weathering rinds and paleosols

There are no studied paleosols in our field locality, although weathering rinds on pebbles do exist and serve a similar purpose [9]. Rinds and broken rinds on felsic sericitized and carbonated volcanic rocks indicate they lingered long enough in transport for chemical alter- ation to occur. This indicates that there was time for chert pebbles to partly dissolve provided the streams were undersaturated in quartz and the kinetics of dis- Fig. 3. Quartz–feldspar–lithic (QFL) ternary plot for Moodies Group solution were rapid. Chert pebbles in the Moodies sandstone, north of the Inyoka Fault and from within the lower Group do not exhibit dissolution features, such as alluvial and fluvial depositional facies. Data collected and compiled dog-tooth edge weathering or micro-karst indentations. by Chandler [11]. Quartz in this diagram includes chert. Most of the Outside of the Barberton Greenstone Belt, we know points lie within the recycled origin tectonic province of Dickinson and Suczek [19]. The low abundance of lithic fragments and feldspar, of one relevant erosion-surface exposure in Australia despite their abundance in the source terrane [11], indicates significant [18]. A chert bed appears to have been resistant to chemical weathering. erosion relative to adjacent granites and greenstones, N.H. Sleep, A.M. Hessler / Earth and Planetary Science Letters 241 (2006) 594–602 597 overlain by a possible paleosol. The chert surface does Angel Falls is a spectacular example of quartzite karst. not show any evidence of dog-toothed or karst features The waterfall exits from a quartzite cave along the cliff that would accompany rapid dissolution in extremely face [28]. Here, karst formed at high elevations at mean hot conditions. temperatures as low as 14 8C and at a very high rainfall rate of ~5 m/yr. It is important to note that development 3. Modern geological analogs of quartzite karst in this region was helped significantly by restricted mechanical weathering along the plateau’s The weathering of modern granitic rocks to quartz- subdued topography. rich sand and clay-rich shale is well known. So is the Chemical analyses of groundwater at Angel Falls put occurrence of some dissolution features in quartz [11]. the karst-forming process into geochemical context. Modern weathering occurs when reaction rates are Effluent water from the local quartzite near Angel fastest, during the wet part of the year in seasonally Falls contains SiO2 at ~2% of saturation [30]. This humid climates and during the warm part of the year in indicates that the dissolution of quartzite is kinetically temperate climates [26]. Modern weathering thus limited. Amorphous silica does occur from local evap- occurs at ~25 8C except at very high latitudes. oration, but plays no part in the overall mass balance as Although lower levels will be tapped during this SiO2 had to dissolve from quartzite in the first intensive erosion events, it is mostly the tops of weath- place. ering profiles that supply grains to fluvial environments Chert also forms a regolith in the warm humid and become the ancient sediments that we can use to mountainous Ashio region of Japan [31]. Water infil- infer surface conditions. Our sand grains encountered trates into the chert regolith and exits from springs. pristine rainwater and far-from-equilibrium conditions Although, dissolution features are not evident within just before mobilization. They preferentially carried the chert bed, some dissolution does occur. Springs records of kinetically limited near-surface processes where just chert is available are at ~29% of quartz away from the outcrop. This is true even if the deeper saturation (Tsuyoshi Hattanji, personal communication, weathering profile was saturated with respect to quartz. 2005). Karst weathering of quartzite to a friable rock also 3.1. Chert and quartzite occurs in the semiarid climate near Delhi, India [32]. Weathering involves strong acids from oxidation of Chert is a relatively uncommon rock in modern pyrite, which is inapplicable to anoxic conditions in source areas. Quartzite is analogous and more abun- the Archean but may be partially analogous to higher dant. Like chert it is nearly monomineralic, has little CO2 levels proposed for the mid-Archean (9). Desicca- porosity, and forms extended outcrops. It differs from tion akin to salt weathering may also be important (J. chert in having a larger grain size, which will play a Donald Rimstidt, personal communication, 2005). Be- role in its mechanical breakdown. Karstic features in yond showing that quartzite can become friable under carbonate-sedimented sandstone are not relevant here semiarid acid weathering in modern climates, this situ- as quartz dissolution is not involved. ation provides little insight into weathering in humid Because chert is such a minor component of most Archean climates. source terranes, its overrepresentation in conglomerate or river gravel is an indication of its resistance to 3.2. Sand from granitic rocks weathering and decomposition to sand-sized fragments, under a range of climatic conditions. Overrepresenta- Quartz dissolution during the weathering of granite tion of chert clasts occurs in relatively recent, humid must be considered separately, due to interference from environments that were warm but not extreme. Creta- the simultaneous dissolution of additional silicates. ceous conglomerates in Iowa provide a example: local- Quartz sand forms from granitic rock when feldspar ly derived chert clasts from Paleozoic carbonates occur and mafic weather to clay. Early in weathering together with , quartzite, and vein quartz derived the dissolution of feldspar may supersaturate the water from the Canadian Shield [27]. with respect to quartz, thus precluding further quartz Quartz, however, does undergo some chemical dissolution. Careful studies show that quartz grains do weathering in modern environments. Modern karst dissolve to some extent in the later stages of modern occurs in quartzite and other hard rocks [28,29],but weathering despite the involvement of feldspar. Schulz is not known to occur on chert (S. Doerr personal and White [33] provide a field study of the modern communication, 2003). The region in Venezuela near weathering of a quartz diorite and include a review of 598 N.H. Sleep, A.M. Hessler / Earth and Planetary Science Letters 241 (2006) 594–602 the literature. Their locality in Puerto Rico receives 4 m expose our ignorance of what a hot Archean climate of rain per year but only 0.6 m soaks in. It takes water would be like. 12 yrs to traverse the 9-m thick regolith profile. The saprolite regolith contains oxides, biotite, quartz, 4.1. Archean climate and landscape and kaolinite. Dissolved silica increases linearly with depth to saturation at the base of the profile. Mass We discuss ~50 and ~75 8C climates in comparison to balance indicates that comparable amounts of SiO2 the modern Earth where weathering occurs at ~25 8C. come from dissolution of biotite and quartz. Latent heat would buffer temperature daily, annually, and Etch pits on grain surfaces are common. Schulz and geographically. At saturation, there would be 0.12 and White [33] calculate that they should form when the 0.39 bars of water vapor in the air, respectively, at 50 and SiO2 concentration is below 10–45% of saturation. This 75 8C. The global average rainfall is of the order of a is a common feature of tropical [34]. Pits form meter per year, as it is limited by the heat from sunlight, quickly, for example, in b8000 yrs on dune sand in like at present. northeastern Australia [35]. Details that would be useful to the field geologist are Dissolution is enhanced because the real surface area not evident without sophisticated meteorological calcu- of the diorite-derived quartz grains is large: ~100 times lations. These include the tendency of rain to fall the geometric (equivalent sphere) area and independent orographically and in local torrential storms, maximum of grain size. Dissolution occurs along grain boundaries wind velocities, and the range of relative humidity. We and microcracks, making the grains somewhat friable, note that halite occurs in the Archean [37], and that it although no macroscopic size reduction of the quartz indicates that the relative humidity was below ~0.75 grains is evident. Schulz and White [33] note that their either locally or temporally. average grain radius is expected to change only from We also have poor constraints on Archean geomor- 0.23 to 0.22 mm in the 100 ka that it remains in the soil phology. Land plants were absent. Conceivably, micro- using kinetic data from Lasaga [36]. bial mats stabilized the surface and enhanced weathering. No land are known from the Archean 4. Inferences about Archean climate [38], but this does not mean that the soils were sterile. The 2.76-Ga Mount Roe (Australia) paleosol shows We wish to compare the above modern analogs with evidence of weathering assisted by microbes [39]. For the Archean record, particularly with the mid-Archean now, we can discount weathering aided by abundant sediments preserved in the Moodies Group. To a first humic acids at 3.2 Ga. order, the Archean sediments we have discussed are We note that a recently denuded and sterilized slope similar to modern ones. If anything, quartz dissolution is not a perfect analog for the Archean. Given time, features are less common in the ancient rocks [11].A however, low-relief surfaces on grade with chemical tempting inference is that the Archean climate was like weathering developed in the Archean, as evidenced the modern. Yet we need to identify observations that by the presence of Archean and mature sands. are particularly temperature sensitive, because there We do not consider places where rapid mechanical were many features of the Archean that uniquely af- denudation on high relief surfaces overwhelmed chem- fected weathering (i.e. elevated CO2 and the absence of ical weathering nor do we consider arid environments. plants). Temperature-sensitive observations include the absence of karst solution features in monomineralic 4.2. Carbonate analog quartz pebbles and on chert erosion surfaces and the survival of abundant polycrystalline quartz and chert in Carbonates form karst as well as microkarst features the sand fraction. Conversely, detection of mild solution on weathering surfaces and pebbles. Carbonate karst features like pits on modern grains in tropical climates has been found in the Archean [40], which implies that would be less useful. (Higher temperatures would only chert or quartzite karst might have been found if they increase their rate of formation.) Such features are were abundant. Typically, the kinetics of carbonate likely to be obscured by low-grade . dissolution is geologically fast. It provides an analog Given the lack of modern hot climates, we apply an where solubility, rather than dissolution kinetics, is rate analogy to quantify the kinetics of quartz dissolution. limiting. Carbonates are common on the modern Earth. Their We discuss quartz solubility to put carbonate solu- dissolution rates are fast enough that solubility often bility in context. Here we are interested in rainwater, limits their denudation rates. Before discussing this, we groundwater, and river water in a humid climate. As N.H. Sleep, A.M. Hessler / Earth and Planetary Science Letters 241 (2006) 594–602 599 abundant humic acids are unlikely in the Archean and grain boundaries dissolve first and the rock disaggre- as there is likely to be significant CO2 in the air, we gates as chert or quartz sand grains, which are thus restrict 3VpHV7 and the ionic strength to b1 mmol. In washed away [28,29]. Chert has more grain boundaries these conditions, the solubility of quartz is a moderate than quartzite, and the smaller grains formed during function of temperature and a weak function of pH and grain-boundary dissolution are more easily carried dissolved CO2. Quartz solubility is 11, 21, 37, and 67 away. Therefore, quartzite and chert karst may develop ppm at 25, 50, 75 and 105 8C, respectively, [41].A at somewhat lower macroscopic dissolution rates than meter per year of saturated infiltration would remove 4, carbonates, where most of the material leaves the sed- 8, 14, 25 m of rock in 1 m.y., respectively. imentary system in solution. The solubility of calcite (CaCO3) in rainwater in the presence of modern air is 50 ppm and its (organic soil) 4.4. Quartz dissolution as a paleothermometer solubility at 0.01 bar CO2 is 160 ppm [42]. The solu- bility is higher if humic acids are abundant, but this To show that quartz dissolution is a paleotherm- situation is unnecessary to form karst, as in the Archean ometer, we consider a 105 8C humid climate. This is [40]. The solubility of dolomite (which also forms higher than we have seen proposed for the Archean, but karst) in rainwater is between that of calcite and quartz we want to start with a case that is observably different at ~25 8C, but is poorly determined [42]. Initial porosity from the modern Earth. We consider a low-relief sur- is not needed for carbonate karst. It forms in hard face with granitic and chert outcrops. By assumption, and marble [e.g., Mueller [43]]. the low-gradient streams in this terrain readily carry material in solution from the watershed, but are inef- 4.3. Quartz dissolution kinetics fective at removing solids including clays. At this temperature, sand size grains equilibrate with Overall, the solubility of quartz is only modestly water on a laboratory scale of weeks [45]. The effluent below that of carbonates. We need to look to the slow water from the chert and quartzite is saturated with dissolution rate of quartz at ~25 8C to explain the rarity quartz. The denudation rate at 1 m per year of rain is of quartzite karst on the modern Earth, especially since 25 m per m.y. effluent water from the Venezuelan quartzite karst has Within a granitic rock in this hypothetical terrane, the only ~2% of quartz saturation [27]. For the Archean, we high temperature speeds up the break-down of need to look at the dependence of dissolution rate on and the dissolution of quartz. The weathered soil is temperature, and we use the far-from-equilibrium rate Al2O3-rich clay according to Wolery [46]. (That is, the in moles per square meter of exposed surface per time. presence of clays alone is not a good paleothermometer.) This dissolution rate of quartz at 50 and 75 8Cisa The clay soil, however, needs to be removed mechani- factor of ~10 and ~100 respectively above the solution cally for open-system weathering to proceed. As already rate at 25 8C and a weak function of pH and CO2 noted, rainfall and runoff were comparable to present, as concentration for the ranges of interest [44]. were mechanical denudation processes that depended on Rainwater falls essentially devoid of SiO2 so that the available water, such as those generating the Moodies bulk of the dissolution occurs at the far-from-equilibrium Group braided stream deposits. The clay surface on rate. Rainwater percolates through the soil dissolving granite becomes the high ground if its denudation rate quartz, leaving the effluent either undersaturated as in is less than that of the chert. This would occur if relief the Venezuelan karst locality [27] or essentially saturated were adequately low, so that seasonal drying might turn as in Puerto Rico [33]. The bulk of the dissolution occurs the clay into a resistant adobe-like surface. at water–quartz contacts where replenishment maintains We now look more closely at the chert and at the far-from-equilibrium conditions. The rest of the real quartz grains housed in the granitic rock. Dissolution grain surfaces are not efficiently replenished, and here along grain boundaries readily disaggregates the chert slow dissolution will occur at or near equilibrium con- and other polycrystalline quartz into a fine-grained ditions. It should be emphasized that field observations material. It also acts along the microcracks in quartz are necessary for calibration as the field rate is less than grains making them friable. At low denudation rates, the laboratory rate and the real surface area of grains is the quartz and chert that do not dissolve end up as clay much greater than equivalent spheres [33]. and silt-sized fragments. Grain size and framework structure of the rock We clearly do not see these features in the Archean. control development of dissolution features related to Chert is highly resistant, underlying high ground and chemical weathering. In a quartzite or chert, the quartz decomposing only partially into pebbles and sand, and 600 N.H. Sleep, A.M. Hessler / Earth and Planetary Science Letters 241 (2006) 594–602 only modest amounts of clay-sized grains. Polycrystal- paleothermometer for surface conditions on the early line quartz is abundant in sands and has not entirely Earth. We concentrated on quartz because it is a simple broken down along fine-scale grain boundaries into silt- common for which chemical data exist. Its or clay-sized fragments. weathering is not strongly affected by atmospheric Can we use these observations, together with con- CO2. The observations are quite low-tech involving straints provided by modern quartz dissolution, to quan- point counts of sediments, visual observation of peb- tify Archean surface temperature? A transition to the bles, and microscopic observations of grains in sedi- expected 105 8C weathering products must occur some- ments and weathered rock. In the Archean as at present, where between that temperature and 25 8C. We expect quartz is a major component of most clastic sediments. that it occurs closer to 25 than 105 8C primarily because Hard monomineralic quartz, as in chert and quartzite, the dissolution rate is an order-of-magnitude function of resisted solution weathering. In the Moodies Group, in temperature over the range of interest, while solubility particular, chert is overrepresented in conglomerate is a factor-of-a-few function. In other words, the tem- relative to the source region. with abundant perature does not have to be much higher than at quartz grains are common. Polycrystalline quartz es- present for kinetically controlled dissolution effects to caped disaggregation. We have no evidence of karstic become evident. features on Archean quartz pebbles or erosion surfaces. We do not have any known climates with weathering It is easier to say that that the Archean data are temperatures significantly above 25 8C, and so we must consistent with modern weathering temperature than extrapolate from measured modern systems. We do note that they preclude elevated weathering temperatures. that the dissolution features of quartz we have discussed Importantly, features related to friability involve posi- do occur modestly on the modern Earth. In particular, tive evidence rather than absence of evidence. In par- faster dissolution at temperatures somewhat above 25 ticular, monocrystalline quartz, polycrystalline, quartz, 8C enhances disaggregation along cracks and grain and chert are common in the Archean sand fraction. boundaries that leads to karstic features. Such dissolu- Our chert pebbles behaved as hard, not friable, rocks. tion involves only a minor volume in the grains so it is As noted in Section 4, friability and disaggregation more sensitive to kinetics than solubility. Its last stages involve small volumes of quartz and hence are mea- (in ancient sediments that we can actually observe) sures of dissolution rates more than solubility. We occurred near the erosion surface by interaction with expect that friability strongly affects the sand fraction far-from-equilibrium rainwater. even at 50 8C. Using the Puerto Rican exposure [33], quartz grains Finally, our conclusions are not in direct conflict in granitic rock became bsomewhat friableQ in ~100 ka of with ocean temperatures inferred from oxygen isotopes exposure time in an area of significant mechanical de- in chert [1]. Our data come from a later time when these nudation although the bulk diameter of the grains marine cherts were exposed to erosion. As noted in remains. That is, the grains are to the point that they Section 4.1, we have no good theoretical constraints would fracture and largely disintegrate during transport. on the long-term and geographic variability of Archean They would break in situ given more weathering or climate. A geologist who discovers pronounced quartz modestly faster dissolution kinetics. Extrapolating line- solution features in ancient sedimentary rocks can ea- arly from the Puerto Rican quartz grains (as dissolution sily turn our argument around. Further examination of rate increases a factor of 10 with every 25 8C temperature the rock record is warranted. increase), the observed amount of friability should occur after 10 and 1 ka at 50 and 75 8C, respectively. Acknowledgements Quartzite karst formed in a favored environment in Venezuela even though the effluent water is ~2% of We thank Don Lowe for help with field and labora- quartz saturation. Again extrapolating linearly, the ef- tory aspects of this study and Dave DesMarais for fluent would be ~20% saturated at 50 8C and nearly discussion of chemical cycles and Archean climate. saturated at 75 8C, thus inferring much more rapid karst Jim Kasting, S. Doerr, Rob Rye, Susan Brantley, Ray formation and quartzite disaggregation. Pierrehumbert, Lee Kump, Marjorie Schulz, Patricia Dove, Don Rimstidt, Tsuyoshi Hattanji, and Roger 5. Conclusions Buick provided prompt responses to our questions. Kevin Zahnle provided helpful comments. Wouter We have applied rather banal observations on Ar- Bleeker and Lee Kump provided helpful comments. chean sediments and modern environments to obtain a Cores were provided by the Sheba Royal Gold Mine N.H. Sleep, A.M. Hessler / Earth and Planetary Science Letters 241 (2006) 594–602 601 in Barberton, South Africa. This work was performed [17] C.E. Heubeck, D.R. Lowe, Depositional and tectonic setting of as part of a collaboration with the NASA Astrobiology the Archean Moodies Group, Barberton Greenstone Belt, South Africa, Precambrian Res. 68 (1994) 257–290. Institute Virtual Planetary Laboratory Lead Team. [18] R. Buick, J.R. Thornett, N.J. McNaughton, J.B. Smith, M.E. Barley, M. Savage, Record of emergent continental crust ~3.5 References billion years ago in the Pilbara craton of Australia, Nature 375 (1995) 574–577. [1] L.P. Knauth, D.R. Lowe, High Archean climatic temperature [19] W.R. Dickinson, C.A. Suczek, Plate tectonics and sandstone inferred from oxygen isotope geochemistry of cherts in the 3.5 composition, Am. Assoc. Pet. Geol. Bull. 63 (1979) 2164–2182. Ga Swaziland group, South Africa, Geol. Soc. Amer. Bull. 115 [20] H. Nesbitt, G. Young, S. McLennan, R. Keays, Effects of (2003) 566–580. chemical weathering and sorting on the petrogenesis of siliclas- [2] P.L. Corcoran, W.U. Mueller, E.H. Chown, Climatic and tec- tic sediments, with implications for studies, J. Geol. tonic influences on fan deltas and wave- to tide-controlled 104 (1996) 525–542. shoreface deposits: evidence from the Archaean Kesharrah For- [21] M. Johnsson, R. Stallard, R. Meade, First-cycle quartz arenites mation, Slave Province, Canada, . Geol. 120 (1998) in the Orinoco River basin, Venezuela and Colombia, J. Geol. 96 125–152. (1988) 263–277. [3] P.L. Corcoran, W.U. Mueller, W.A. Padgham, Influence of [22] J.A. Donaldson, E.A. de Kemp, Archaean quartz arenites in the tectonism and climate on lithofacies distribution and sandstone Canadian shield: examples from the Superior and Churchill and conglomerate composition in the Archean Beaulieu Rapids Provinces, Sediment. Geol. 120 (1998) 153–176. Formation, Northwest Territories, Canada, Precambrian Res. 94 [23] W.U. Mueller, P.L. Corcoron, Late-orogenic basins in the Ar- (1999) 175–204. chaean Superior Province, Canada: characteristics and infer- [4] G.M. Young, V. von Brunn, D.J.C. Gold, W.E.L. Minter, Earth’s ences, Sediment. Geol. 120 (1998) 177–203. oldest reported glaciation: physical and chemical evidence from [24] K.N. Sircombe, W. Bleeker, R.A. Stern, Detrital geochro- the Archean Mozaan Group (similar to 2.9 Ga) of South Africa, nology and grain-size analysis of a ~2800 Ma Mesoarchean J. Geol. 106 (1998) 523–538. proto-cratonic cover succession, Slave Province. Canada, Earth [5] J.K. Kasting, Earth’s early atmosphere, Science 234 (1993) Planet. Sci. Lett. 189 (2001) 207–220. 1383–1385. [25] B.G. Els, The auriferous Archaean systems of [6] J.F. Kasting, J.L. Siefert, Life and the evolution of earth’s South Africa: unique palaeo-environmental conditions? Sedi- atmosphere, Science 296 (2002) 1066–1068. ment. Geol. 120 (1998) 205–224. [7] A.A. Pavlov, J.F. Kasting, L.L. Brown, K.A. Rages, R. Freed- [26] D.D. Ekart, T.E. Cerling, I.P. Montan˜ez, N.J. Tabor, A 400

man, Greenhouse warming by CH4 in the atmosphere of early million year carbon isotopic record of pedogenic carbonate: earth, J. Geophys. Res. 105 (2000) 11981–11990. implications for paleoatmospheric carbon dioxide, Am. J. Sci. [8] N.H. Sleep, K. Zahnle, Carbon dioxide cycling and implications 299 (1999) 805–827. for climate on the ancient earth, J. Geophys. Res. 106 (2001) [27] R.L. Brenner, G.A. Ludvigson, B.L Witzke, P.L. Phillips, T.S. 1373–1399. White, D.F Ufnar, L.A. Gonza´lez, R.M. Joeckel, A. Goettemoe- [9] A.M. Hessler, D.R. Lowe, R.L. Jones, D.K. Bird, A lower limit lerr, B.R. Shirk, Aggradation of gravels in tidally influenced for atmospheric carbon dioxide 3.2. billion years ago, Nature fluvial systems: Upper Albian (Lower Cretaceous) on the cra- 428 (2004) 736–738. tonic margin of the North American Western Interior foreland [10] R. Rye, P.H. Kuo, H.D. Holland, Atmospheric carbon-dioxide basin, Cretac. Res. 24 (2003) 439–448. concentrations before 2.2-billion years ago, Nature 378 (1995) [28] S.H. Doerr, Karst-like landforms and hydrology in quartzites of 603–605. the Venezuelan shield: pseudokarst or brealQ karst, Z. Geomor- [11] F.W. Chandler, Quartz arenites: review and interpretation, Sed- phol. 43 (1999) 1–17. iment. Geol. 58 (1988) 105–126. [29] R.A.L. Wray, A global review of solution weathering forms on [12] A. M. Hessler, Evidence for climate and weathering in siliclastic quartz sandstones, Earth-Sci. Rev. 42 (1997) 137–160.

sedimentary rocks of the 3.2 Ga Moodies Group, Barberton [30] M. Mecchia, L. Piccini, Hydrogeology of SiO2 geochemistry of Greenstone Belt, South Africa, Ph.D. Stanford University, the Aonda Cave system (Auyan-Tepui, Bolivar, Venezuela), 2001, p. 258. Boll. Soc. Venezolana Espel. 33 (1999) 1–11. [13] K.A. Eriksson, Alluvial and destructive beach facies from the [31] T. Hattanji, Y. Onda, Coupling of runoff processes and sediment Archean Moodies Group, Barberton Mountain Land, South transport in mountainous watersheds underlain by different sed- Africa and Swaziland, Fluvial , Canadian Society imentary rocks, Hydrol. Process. 18 (2004) 623–636. of Petroleum Geologists Memoir, vol. 5, 1978, pp. 287–311. [32] J.K. Tripathi, V. Rajamani, Weathering control over geomor- [14] K.A. Eriksson, Marginal marine depositional processes from the phology of supermature Proterozoic Delhi quartzites of India, Archean Moddies and Fig Tree Groups, Barberton Mountain Earth Surf. Process. Landf. 28 (2003) 1379–1387. Land, South Africa: evidence and significance, Precambrian [33] M.S. Schulz, A.F. White, Chemical weathering in a tropical Res. 8 (1979) 153–182. watershed, Luquillo Mountains, Puerto Rico III: quartz dissolu- [15] K.A. Eriksson, Marginal marine depositional processes from the tion rates, Geochim. Cosmochim. Acta 63 (1999) 337–350. Archean Moddies and Fig Tree Groups, Barberton Mountain [34] H. Eswaran, G. Stoops, Surface textures of quartz in tropical Land, South Africa: evidence favoring an Archean continental soils, Soil Sci. Soc. Am. J. 43 (2) (1979) 420–424. margin, Precambrian Res. 12 (1980) 141–160. [35] K. Pye, J. Mazzulo, Effects of tropical weathering on quartz [16] Heubeck, C.E. Geology of the Archean Moodies Group, central grain shape: an example from northeastern Australia, J. Barberton Greenstone Belt, South Africa. Ph.D. thesis, Stanford Sediment. Res., Sect. A Sediment. Pet. Proc. 64 (3) (1994) University, 1993, 390 pp. 500–507. 602 N.H. Sleep, A.M. Hessler / Earth and Planetary Science Letters 241 (2006) 594–602

[36] A.C. Lasaga, Chemical kinetics of water–rock interactions, J. [42] K.B. Krauskopf, D.K. Bird, Introduction to Geochemistry, Geophys. Res. 89 (1984) 4009–4025. McGraw-Hill, New York, 1995, 647 pp. [37] C.A. Boulter, J.E. Glover, Chert with relict hopper molds from [43] M. Mueller, Karst hydrogeology of the Takaka Valley, Golden Rocklea Dome, Pilbara Craton, western Australia — an Archean Bay, Northwest Nelson, N. Z. J. Geol. Geophys. 34 (1991) halite-bearing evaporite, Geology 14 (1986) 128–131. 11–16. [38] R. Buick, Life in the Archean, in: D.E.G. Briggs, P.R. [44] P.M. Dove, The dissolution kinematics of quartz in sodium- Crowther (Eds.), Paleobiology II, Blackwell Science, Oxford, chloride solutions at 25-degrees-C to 300-degrees-C, Am. J. 2001, pp. 13–21. Sci. 294 (1994) 665–712. [39] A. Neaman, J. Chorover, S.L. Brantley, Element mobility pat- [45] H.U. Rimstidt, H.L Barnes, The kinetics of silica–water reac- terns record organic ligands in soils on early earth, Geology 33 tions, Geochim. Cosmochim. Acta 44 (1980) 1683–1700. (2005) 117–120. [46] Wolery, T. J. EQ3/6: a software package for geochemical [40] T.P. Kusky, P.J. Hudleston, Growth and demise of an Archean modeling of aqueous systems. (The Regents of the University carbonate platform, Steep Rock Lake, Ontario, Canada, Can. J. of California and Lawrence Livermore National Laboratory, Earth Sci. 36 (1999) 565–584. 1997). [41] J.D. Rimstidt, Quartz solubility at low temperatures, Geochim. Cosmochim. Acta 61 (1997) 2553–2558.